Years ago, Fisher and Allen (7) suggested that the characteristic
damage observed with P. aeruginosa eye infections
may be attributable to the action of bacterial proteases on
the corneal proteoglycan. Several groups, including Kawaharajo
et al. (16), Hirao and Homma (12), Kreger and Gray
(20) and Kreger and Griffin (21) have since provided data to
support this hypothesis. Recently, Ohman et al. (27) initiated
Isolation of mutants specifically deficient in the production
of alkaline protease would facilitate definitive studies
addressing the role of alkaline protease in P. aeruginosa
infections. This approach has been complicated by the fact
that most strains of P. aeruginosa produce at least two
distinct extracellular proteases (25) and that alkaline protease
does not possess a unique or stringent substrate specificity
(26). We have been unable to specifically inhibit the
activity of interfering protease (elastase) to facilitate detection
of alkaline protease mutants on skim milk agar plates
(unpublished data). To circumvent these problems, we chose
to isolate alkaline protease mutants in strain PA103, which
produces alkaline protease (4) but no detectable elastase
(28). Despite the fact that strain PA103 is serum sensitive
(41), it is virulent in a mouse eye model of corneal infection
(17, 27). We have isolated alkaline protease-deficient mutants
of strain PA103 after ethyl methanesulfonate (EMS)
mutagenesis and screening of mutagenized colonies on skim
milk agar plates. After in vitro characterization of these